1

Godin Tepe: The Inscriptions

by

William W. Hallo

Yale University

I. Introduction: the discovery of the tablets and my connection

Thirty years ago - in another era! - Cuyler Young

circulated a memo to "all the principal authors for the Godin

Project final publication." 1 My contribution was foreseen for

Volume IV: "The Early Periods at Godin Tepe," more specifically

for its Fascicle A: "Period V," and within that presumably for

Part 3: "Small finds." Material for the fascicle, it was

optimistically suggested, was to be turned in by June 1976.

As fate would have it, this deadline was not met, to

say the least. In my own case, a ceaseless avalanche of other

obligations competed for prior attention. The reasons that I am

now at last able to move the project to the front burner are

twofold. One is the luxury of retirement (as of July 1, 2002), the

other is Cuyler's patience, rivalling that attributed to Job by

James.

It may well be asked why I was assigned to deal with

the tablets in the first place. The answer begins with a chance

meeting between Cuyler and me - in a New York subway if memory

2

serves - early during my curatorship of the Yale Babylonian

Collection, which I had taken over in 1963. It eventually led to

my proposal, and his agreement, that the Collection make a modest

financial contribution to his excavation at Godin Tepe, which had

been going on since 1965. In return our faculty and students could

participate, as indeed happened, with Richard Ellis and his wife

Maria deJong Ellis, Harvey Weiss, Marie-Henriette Carr and her

husband Charles (Charlie) Gates, and Curtiss Hoffman all involved

at one time or another.

"The fifth and final season of excavations at Godin

Tepe began on May 19 and ended on August 30, 1973" (Young and

Weiss 1974:207). On June 6, Charlie excavated the first inscribed

text at Godin, an event dramatically recounted by Cuyler in the

magazine of the Royal Ontario Museum under the title "The Day We

Found the Tablet" (Young 1974). My own visit to the site took

place less than two months later. Cuyler assured me I'd have no

trouble finding it, since I had told him I was bringing Edith with

me. It was not, however, Edith Porada as he thought, but my late

wife.

Our ride from Teheran via Hamadan to Kermanshah was an

adventure. The road had been improved two years earlier in

connection with the Shah's 2500-year jubilee, but not so the

Mercedes limousine which we rented for the occasion. The driver

held steadfastly to the middle of the road, but it was his tires

that troubled us most. Their tread was 98% gone. When one of them

3

lost the remaining 2%, it was providentially in front of an auto

mechanic's shop. The "spare" was if anything more parlous than all

the others, but the young mechanic patched it well enough for it

to see us to our destination. And thus I earned the chance to edit

the Godin tablets! (See Weiss and Young 1975:88, n. 14; Hallo

1980:309 and n. *.)

In January, 1979, Cuyler Young provided me with

photographs of the inscribed sides of 27 of the tablets, i.e., all

those reasonably well preserved, and in May of the same year

followed that up, at my request, with "xerox copies of the

drawings on the registration cards of all thirty-six tablets and

tablet fragments which had marking of some sort (including finger

nail impressions)," to quote his covering letter of May 9, 1979.

The drawings were prepared by L. R. [Cuyler: PLEASE SUPPLY FULL

NAME], and included 25 of those I had in photograph form; 11

others I have only in drawings and 2 only in photographs (see

below, IV). From that day to this (2003), however, I have not

looked at either the drawings or the photographs. As for the

original tablets, I have not seen them since my visit to Godin in

1973. I have kept abreast of much of the literature on the

forerunners of writing in the ancient Near East, and have taken up

my long delayed assignment as soon as possible after my retirement

from active teaching.

Of course there can hardly be a question of a standard

text publication for these 38 tablets. (The number 43 quoted in

4

early publications includes fragments that had neither

inscriptions nor seal impressions.) Instead, they will be

presented here in terms of their wider context. Where do they fit

into the pattern of other archaic Near Eastern writing, the

evolution of literacy and "numeracy," and the cultural horizon of

"Greater Mesopotamia"? As a philologist I am forced to ignore all

strictly archaeological questions, but happily can refer the

reader to a dissertation on the subject by one of Cuyler's

students (Badler 19XX).

II. The parallels from elsewhere

The tablets from Godin all come from Level V, now dated

by the excavator to ca. 3500-3200 BCE (Young 1997; previously

Young 1986:222). That means they are as early as or even slightly

earlier than the earliest texts from Uruk, which come from Level

IVa, a level now dated, partly on the basis of Godin, to ca. 3200-

3100 (Porada 1965:177; Porada et al. 1992:I 100, II 96). But the

Uruk material is not yet suitable for comparative dating purposes.

As van Driel put it, "no tablets have been published (or found)

from a context earlier than Uruk IVa, but on the other hand, the

suspicion remains that the tablets are used to date their

archaeological context to Uruk IVa. There is, however, as yet no

reason to assume that 'writing' was unknown before the buildings

regarded as Uruk IVa was in use.... Uruk, though eponymous for the

5

culture, is unsuitable for serious comparative dating purposes"

(1982:18).

In any case, the Godin tablets are not unique. Similar

tablets have been found in a wide arc of sites around the Near

East (see the map) and even beyond it. They will be surveyed

briefly here.

MAP copied from Schmandt-Besserat 1992 I 131.

To begin with Uruk itself, there are a few Godin-type

tablets, i.e. with numerals only to the exclusion of word-signs.

To those listed by van Driel (1982:18, n. 6) and Schmandt-Besserat

(1992:130, nn. 25-31), may now be added one which has been in the

British Museum since 1851 (Reade 1992) and another that has just

been published from the Heidelberg collection (Englund and Nissen

2001:W 20239; cf. also W 19410,11; W 19689,a, W 20044,11). As to

the more fully developed script found in Level IVa, the forerunner

of true cuneiform sometimes referred to as proto-cuneiform or

Proto-Keilschrift (Vaiman 1976:15), it includes numerical signs

very similar to those in use at Godin. The majority of the texts

in this script may be described as accounts; numerals figure

prominently in them and often help in their decipherment. Great

strides have been made in their interpretation since the first

systematic attempts by Falkenstein (1936), Vaiman (1976) and

6

others, especially by Nissen, Damerow and Englund (1990, 1993).

When we turn to Iran itself, the comparative material

is more abundant. At Choga Mish in Khuzistan, some 20 miles from

Susa, were found a few "sealed clay balls and crude clay tablets

with numbers only, but no real writing" (Nissen 1977:19, citing

Delougaz 1972, p. 27, 30 and pls. 9f.), exactly as at Godin. But

Susa itself has yielded a rich harvest of bullae and Godin-type

tablets. Here they represent the first stage in an evolution that

led by the addition of the first pictograms in Level 16 (LeBrun

and Vallat 1978; Vallat 1979) ultimately to the fully developed

Proto-Elamite script (see e.g. Stolper 1978:94-96; below, VII A).

Other Iranian sites that have yielded archaic tablets are Tepe

Sialk, though these are more properly described as proto-Elamite

(Ghirshman 1938:65-68 and pls. xciif.), and Tall-i-Ghazir (Weiss

and Young 1975:91 and n. 20).

The situation in northern Mesopotamia is similar. From

Nineveh in the East (Collon and Reade 1983:33f.), via Nagar/Nawar

(= Tel Brak, Oates 1982:65; Matthews and Eidem 1993; Oates, Oates

and McDonald 2001), to Habuba Kabira (Strommenger 1977, 1981) and

Jebel Aruda (Van der Leeuw 1974:fac. p. 83; Van Driel 1982) in the

West and back to Mari (Parrot 1965:12) and Khafaje (Tutub; Weiss

and Young 1975:88 n. 16) there have been finds of archaic number

tablets and in some cases such as Habuba Kabira (Heinrich et al.,

1973:26; Strommenger 1973:170) also bullae or, at Brak, "dockets"

(cf. Matthews 1997, Nos. 117, 164, 201).

7

Going even further afield, the analogies are less

clear. Shahr-i Sokhta, which may (Hansman 1978) or may not

(Majidzadeh 1976; Michalowski 1986:133) be ancient Aratta, has

yielded one tablet with both number-sign and word-sign and is

described as Proto-Elamite by Meriggi (apud Amiet and Tosi 1978:28

and fig. 16). But it is difficult to judge a newspaper account of

Pakistani finds of the earliest writing from the Indian

subcontinent, allegedly dating to 3300 BCE (Anon. 1977). And it is

highly unlikely that there is any connection with the so-called

Tartaria tablets from Romania (Vlassa 1963), let alone the pre-

neolithic signs from Russia (Marshack 1979).

III. The implications for the origins of writing

In a 1980 survey of cuneiform writing, Edzard weighed

the possibility of an antecedent system based on three-dimensional

clay "tokens" while withholding outright assent to the theories of

Schmandt-Besserat (Edzard 1980:548f.). My own endorsement was less

qualified, though I recognized the objections raised against the

thesis (Hallo apud Schmandt-Besserat 1992 I:ix-xi; 1996:26-30).

Since then, the ground has been gone over again by Glassner, who

took note of my position (2000:90) and even claimed to have found

the Sumerian term for the tokens (or counters or calculi or

stones) which form the basis of Schmandt-Besserat's thesis.

According to him it is imna (2000:103), although this is not

8

equated with abnu, the Akkadian term which designated the

phenomenon in a later Nuzi text.

Summing up my own reading of the thesis, writing was

preceded by counting, and counting was done with clay tokens such

as occur as early as the ninth millennium BCE throughout the Near

East, i.e., shortly after the neolithic revolution or

"agricultural revolution" and probably as a consequence of it.

After some millennia of simple token assemblages (1), it was found

convenient, before the end of the fourth millennium, to string

the tokens together and enclose the end of the string in a ball of

clay ("bulla") or to deposit them inside round and hollow clay

envelopes (2). These bullae and envelopes were impressed with

stamp seals or cylinder seals and, in the case of some envelopes,

with tokens like those enclosed in them to indicate what they

contained. But their format was not ideal for record-keeping. To

verify the contents and to reuse the tokens required breaking the

envelope open. It was simpler to rely on the impression of the

tokens on the outside of the envelope and simpler still to

dispense with the making and enclosing of ever new tokens and to

rely exclusively on their impressions on the outside, or even to

imitate them with stylus or finger (so Nissen, Damerow and Englund

1990:171, fig. 18c, 173 = 1993:127, fig. 109) or fingernail (3).

That given, it was only a short logical step to abandoning the

envelope shape entirely in favor of a simple rectangular tablet

whose shape was only slightly rounded on the writing surface (4).

9

In short order the rounding of the obverse writing surface was

replaced by a flat obverse, presumably in order to prevent the

signs of the obverse from being deleted by contact with the palm

of the hand while inscribing the reverse (5). The final

transformation occurred when a reed stylus was employed to impress

the clay tablet with designs resembling in two-dimensional format

the three-dimensional tokens that had preceded (6). With this

step, full writing had been achieved (Hallo 1996:27f.). Much the

same sequence of steps had already been outlined in 1978 by LeBrun

and Vallat.

Where do the Godin tablets fit into this sequence? In

my terms, they belong to step (4) above, when the envelope shape

had been abandoned in favor of rectangular tablets slightly

rounded on both obverse and reverse, but before the flattening of

the obverse which had become standard. In Schmandt-Besserat's own

terms, they are coeval with Uruk IVa and predate Uruk III; they

thus anticipate or perhaps coexist with the earliest forms of

pictographic writing at Uruk (1992:133).

IV. Physical Description of the Godin Tablets

A. Catalogue of texts

In the following catalogue, the tablets have been

renumbered from 1-38, retaining the sequence of field numbers

assigned to them in 1973. Tablets of which photographs were

10

available to me are indicated by P, those for which drawings were

available by D followed by the number assigned to the drawing on

the registration card. Dimensions are given, in centimeters, as on

the registration cards (where available) as follows: a = height of

the tablet measured from the top as "top" is interpreted below (V

A); b = width of tablet; c = maximum thickness of tablet. Plus (+)

indicates that the relevant dimension of a fragmentary tablet was

originally greater. Sealed tablets are indicated by S, unsealed

ones by U. Some of the tablets have been previously published by

Weiss and Young (1975) and/or by Schmandt-Besserat (1981, 1992);

they are identified here by Y/W, W/Y and S-B respectively.

No. Gd no. P/D a b c Seals Previous

publication

1. Gd 73-53 D 17 3.2+ 2.3+ 1.3 ?

2. Gd 73-54 P/D 16 3.4 3.5+ 2.1 S W/Y 89 (1)

3. Gd 73-61 P/D 14 3.4 5.3 1.8 U

4. Gd 73-64 P/D 27 S W/Y 89 (5)

5. Gd 73-153 P/D 40 3.95 5.55 2.15 S W/Y 90 (5)

S-B 1992:132,

fig. 80

6. Gd 73-154 P/D 22 4.15 3.7+ 2.1 U

7. Gd 73-155 P ?

8. Gd 73-156 P/D 29 2.4+ 3.4+ 1.85 U?

9. Gd 73-157 D 23 4.9 5.3 1.85 U

11

10. Gd 73-158 P/D 24 4.0 5.75 2.25 U

11. Gd 73-159 P/D 25 4.3 5.5 1.9 U

12. Gd 73-160 P U

13. Gd 73-161 P/D 28 4.65 4.15+ 2.15 S

14. Gd 73-162 P/D 39 3.6 5.4 2 U

15. Gd 73-250 P/D 164 5.0 3.8+ 2.7 U

16. Gd 73-274 P/D 172 3.65 4.85 1.6 U

17. GD 73-279 D 1032 2.7+ 2.3+ 1.7 U

18. Gd 73-286 P/D 1043 3.25 4.7 1.9 U W/Y 89 (6);

S-B 1992:134:fig.81.

19. Gd 73-287 P/D 161 3.1+ 4.2+ 2 U

20. GD 73-288 D 1037 3.9 5.3 2.1 S

21. Gd 73-289 P/D 1036 3.9 5.55 2.15 U

22. Gd 73-290 P/D 1041 2.7 3.9 1.8 U W/Y 89 (3); S-B

1992:136, fig. 89

23. Gd 73-291 P/D 1040 3.5 4.9 1.9 U S-B 1981; 1992:135,

fig. 84

24. Gd 73-292 P/D 1028 3.9 5.5 2.6 U? W/B 90 (1); S-B

1981 fig. 1 ("obv."; 1992:136, fig. 90 ("rev.")

25. Gd 73-293 P/D 166 4.25 4.9 1.9 U

26. Gd 73-294 D 1044 4.35 5.95 2.3 U S-B 1993:136, fig.

92

27. Gd 73-295 P/D 532 3.9 5.95 2.25 S W/Y 89 (2)

28. Gd 73-296 D 185 4.15 5.0 1.5 U?

29. Gd 73-297 D 535 3.4 4.9 2.0 U

12

30. Gd 73-298 P/D 1034 3.55 5.25 1.75 U

31. Gd 73-299 P/D 1039 3.8 5.2 2.2 U

32. Gd 73-317 P/D 1051 3.7 3+ 2.7 U

33. Gd 73-318 P/D 203 4.65 5.6+ 2.15 U

34. Gd 73-319 D 1049 3.8 5.1 1.95 U

35. Gd 73-320 D 167 3.8 5.1 1.9 S W/Y 89 (6)

36. Gd 73-321 P/D 1038 3.8 4.8 2 U

37. Gd 73-324 D 205 3.7 2.7+ 1.9 U

38. Gd 73-415 D 551 2.65+3.3+ 2.7 S

B. Dimensions

It may be noted that the dimensions of the tablets are

remarkably uniform. Apart from fragmentary ones, their height

varies from 2.7 cm (No. 22) to 5.0 cm (No. 15), with the rest

ranging between 3.4 and 4.9. Their width varies from 3.9 (again

No. 22) to 5.95 (Nos. 26 and 27), with the rest ranging between

4.7 and 5.75. Their thickness varies from 1.3 (No. 1) to 2.7 (Nos.

15, 32 and 38) with the rest ranging from 1.5 to 2.6. According to

Weiss and Young (1975:88), their arithmetic mean dimensions are

3.8x5.2x1.9 cm.

C. Seal impressions

Most of the tablets are unsealed. Of the seven (CHECK)

that definitely are sealed, three (2, 20, and perhaps 24) are

"duplicated" by three unsealed tablets (see VI B below).

To the perennial question of which came first, the seal

13

impression or the impressed signs, the answer seems to be that the

seals were impressed first, since the number signs occasionally

seem placed so as to avoid disturbing the seal design as much as

possible. Occasionally the tendency to put the sign or signs in

the middle of the (top) line seems dictated by the prior placement

of the seal impression, as illustrated best by examples from Uruk

such as W 10133 (UVB 3 pl. 19b = 5 pl. 23c) or W16184 (UVB 8 pl.

51c).

14

The same conclusion was reached for the single tablet

from archaic Niniveh (Collon and Reade 1983:33) as well as for

archaic tablets from Uruk and Shuruppak (Matthews XXXXX), while

the opposite is more often true of later periods, e.g. Ur III

(Fischer 1997:98-100; ref. courtesy R. Mayr).

D. Direction of script etc.

On the basis of No. 27, one may read all the Godin

tablets as if they were to be held with the top of the vessel at

the top of the tablet, in keeping with comparable tablets from

other sites. Their signs thus begin at the top of the tablet and

either begin at the right end or are placed in the middle. If a

second line is needed, it either follows immediately after the

first line or, more rarely, is placed at the bottom of the tablet.

With all these conventions may be compared the earliest tablets

from Susa (Scheil 1923; Amiet 1972:474f., 521, 534, 606, 621!,

629, 633, 652, 657, 666, 673; Vallat 1971:fig. 43:9, pl. xxiv 5;

1973:103:4) and a tablet from Habuba Kabira (Heinrich et al.

1973:25 fig. 4). But there are deviations from these patterns,

perhaps for the seal impressions (e.g. Amiet 1972:491, 642).

There seems to be a strong sense of symmetry at work in

the disposition of the signs (cf. Schmandt-Besserat 1992:135),

leading not only to their placement in the middle of the tablet

but to the alignment of signs on the lower lines beneath those on

the higher one. This is noticeable in some tablets from Godin

(e.g. Nos. 18 and 22) and others from Uruk (especially Reade

15

1992). Sometimes the effect is almost pictorial, as at Brak (Oates

1982:65 fig. 51), Habuba Kabira (Strommenger 1973:171), and Jebel

Aruda (van Driel 1982, figs. 1a, b, 4,5) and one has to wonder

whether that is not their primary motive! It certainly strengthens

the case for reading the lines from top to bottom not from left to

right.

If there is not enough room for all the signs of a

given order of magnitude on one line, they are carried forward to

a second line which is begun, unlike the first, at the left edge

(Nos. 10, 24 and 26).

E. Obverse and reverse

The Mesopotamian convention of flattening the obverse

to create what may be called the "plano-convex" tablet developed

early; for a possible motive behind this development see above

(III). But this convention had not yet been fully adopted in the

time of the Godin tablets, most of which are still slightly convex

on both sides; two are flat on both sides (Nos. XXX and XXX) and

only four are plano-convex Nos. XXXXXXXXXX CUYLER PLEASE

SPECIFY!!; cf. Weiss and Young 1975:88). Except for the last,

there are thus no easy criteria for distinguishing between obverse

and reverse. The designations here employed are to that extent

arbitrary, as indicated by the quotation marks; they follow those

supplied with the photographs.

By way of comparison, it may be noted that at

Jebel Aruda, the tablets are variously described as flat (2), flat

16

on one side and convex on the other (4, 10, 11), flat and round

(5), both sides slightly curving (i.e., convex) (7), uninscribed

side slightly flatter than the inscribed side (8, 9), both sides

slightly concave (13). Sometimes the flat side is inscribed (2, 4,

10) and sometimes the convex side is sealed (4, 12?) but at other

times the convex side bears the inscription (11??) and sometimes

the flatter side has the seal impression (9). Since none of the

tablets are inscribed on both sides, the motive for a flat

"obverse" (above) is absent; if in fact the seal was impressed

first, one would rather expect it to appear on the flat side.

Clearly there is as yet no conventional usage in this respect at

Jebel Aruda, and obverse and reverse are arbitrary designations.

A further question is whether the direction of writing

is the same on obverse and reverse or, as in later Mesopotamian

usage, the opposite. In "proto-cuneiform" and proto-Elamite, both

conventions are attested (Vaiman 1976:16) and a similar variety

prevails at Godin. Some of the Godin tablets inscribe both sides

in the same direction (Nos. 3, 14, and 36); one of these tablets

(No. 3) has margins in addition (see IVC above). In Nos. 4, 16(?)

and 31, where the inscribed surfaces are not obverse and reverse

but "obv." and edge, the evidence is less clear. But in Nos. 18,

20, 23, 24? and 26 the later Mesopotamian rule applies. (Cuyler:

THIS NEEDS TO BE CHECKED AGAINST THE ORIGINALS!)

F. Edges.

While most of the tablets use obverse and/or reverse as

17

writing surface, a number of them employ one of the long edges

either in addition to (Nos. 4 and 31) or instead of (No. 5) the

other surface(s). In No. 16, it is not clear whether the deep

impressions on the short edge are intended as signs.

G. "Margins" and "cases"

Two of the tablets, Nos. 3 and 18, have indentations

running all around both obverse and reverse, as if to form a frame

or margin for the signs. No. 14 has such a frame but only on the

"obverse." No. 18 is said to also have a dividing line creating

the effect of two cases (see below, V B 8).

V. Interpretation of the texts.

When it comes to the interpretation of the texts from

Godin Tepe, it is hard to improve on the treatments of Weiss and

Young (1975) and Schmandt-Besserat (1992:129-154). But while they

were able to present the best preserved tablets in their sampling,

the attempt here will be at an exhaustive presentation of all 38

inscribed tablets. The texts may be divided into three unequal

groups, the first consisting of one tablet with an incised drawing

of a pictographic sign as well impressed signs, the second of 25

tablets with impressed signs only, and the third of 11 fragments

to poorly preserved to allow for even a tentative interpretation.

The seal impressions of all three groups, are dealt with elsewhere

in this volume. (CUYLER: YES?)

18

A. The tablet with the pictographic sign (No. 27).

The pictographic sign on No. 27 almost certainly

represents a pitcher or other clay vessel, with pointed base and a

neck secondarily attached to the body. The vessel pictured is of a

type equally familiar from the excavations and from early sign-

forms, both conveniently catalogued by Salonen (1966). For the

former, one may compare especially his plate xli, described as a

"crouching figure with vessel, of limestone, from Susa, archaic

deposit of Stratum Cc of the high proto-Elamite period, third

quarter of the fourth millennium" (Salonen 1966: 481; my

translation); this accords well with both the period and the

provenience of the tablet. For the latter, one may compare his

sign no. 15.1, based on Labat 1948:88f., no. 108), the

pictographic forerunner of the later DUR-sign, itself a ligature

of a sign for a vessel with an inscribed sign for a carrying-cord,

and based on evidence from Jemdet Nasr, i.e. the Uruk III period

(Salonen 1966:37 and 41, based on Deimel 1928:281). To this one

may add the evidence of the Uruk IV period at Uruk itself, where

signs 156-158 in the older sign-list (Falkenstein 1936) and signs

88b and 296 in the newer one (Green and Nissen 1987) closely

resemble ours.

The presence of the pictographic sign allows us to

"orient" this tablet and presumably all the Godin tablets. Vessels

with pointed bottoms clearly could not stand on their own; they

were therefore placed on stands such as those pictured in archaic

19

(Uruk IV and III) signs 161-164. 170 and 178-179 (Falkenstein

1936:45-47) and in slightly later script by the pictographic

ancestor of the sign GAN (Labat 1976:100f. No. 143). Alternatively

they could be "hung" from a cover(?) like that pictured by archaic

sign 181 (Falkenstein 1936:49) and its later equivalent SAKAN (=

GAN+U). All these signs became logograms for various specialized

kinds of vessels, while the standard logogram for vessel, DUG,

derived from the pictogram of a jar, sometimes with the addition

of a spout or handle.

The pictographic sign on our tablet is thus safely

interpreted as a vessel standing on its pointed bottom, and allows

us to "read" the tablet from top to bottom and to describe the

three elongated signs to its left as "verticals." These verticals

are topped by three circular signs above them, but in such a

fashion that the first circular sign is to the right of the first

vertical and the last vertical is to the left of the last circle.

The direction of writing in Mesopotamia was from top to bottom

within the case and from right to left between cases throughout

the fourth and third millennia and did not change to the later

convention of writing from left to right within the case and from

top to bottom between cases until well into the second, especially

for monumental texts (Edzard 1980:545f.; Hallo 1982:114f.). We can

therefore safely deduce that the pictographic sign was written

first, the circles next, and the verticals last. This

sequence provides an important clue to the meaning of the

20

"numerical" signs, since it seems likely if not incontrovertible

that the higher magnitude was written before the lower. On the

analogy of later texts from Mesopotamia, we may tentatively assign

to the circles the meaning "10" and to the verticals the meaning

"1." We thus arrive at the sense "vessel containing 33 unities of

...." or "33 vessels containing ...."

A further observation is in order. The circles and

verticals were impressed into the wet clay by means of tokens or,

with Weiss and Young (1975:88), of a stylus, while the

pictographic sign was definitely drawn on it with some kind of

stylus or other writing instrument. This implies the absence of a

suitable token to depict the vessel. Such tokens have been

identified by Schmandt-Besserat from a large variety of sites, not

including Godin (1992a xxxi and xxxvi sub 13). But since verticals

and circles, or what she calls cones and spheres, were likewise

lacking from the Godin assemblage, it seems more reasonable to

attribute the resort to a drawing of the vessel to the fact that

the tablet represents a step in the transition from impressed

tokens to incised pictograms.

For the possibility of a second tablet with an incised

sign see below, ad no, 23.

B. Whole tablets with impressed signs and no

pictograms.

There are 37 tablets with impressed signs on them but

no pictograms. 25 of these are in reasonably good state of

21

preservation and may be presented here in descending order of the

amounts recorded on them or on their principal surface. For the

relative and absolute values here assigned to the signs, see below

(VII). Whether these amounts refer to the same commodity as the

tablet with the incised pictogram is not clear, but that pictogram

itself reveals little or nothing about the commodity involved, so

would be of no great help in answering the question (but see

below, VI A).

1. 2 large verticals over 4 circles over 6 verticals (=

246?)

No. 33: in every case, the tablets are inscribed

beginning near the top of the writing surface. (For a similar

large vertical see No. 37 (fragment 10) and Amiet 1972:545).

Unless that rule is violated here, the two deep impressions over

the circles must represent a line of writing. In later cuneiform,

verticals most often stand for powers of 60. If the smaller

vertical here too stands for sixty to the zero power = 1 (below,

No. XXX), then the large verticals would likely be sixty to the

power one = 60. But that can hardly be the case, given that two

tablets (below, 2) appear to express 90 by 9 circles rather than

by a sign for 60 and three circles. So it seems safer for now to

suggest a ratio of 1:10 as we move from circles to large

verticals, i.e., to interpret the latter as standing for 100. The

same ratio appears to apply to as we move down from circles to

(small) verticals, based (a) on the analogy of later Mesopotamian

22

cuneiform and (b) on archaic tablets from Susa with nine but never

more than nine impressed verticals(?!) (Scheil 1923:82 = Amiet

1972:520; LeBrun and Vallat 1978:47 fig. 4:5 = 5 pl. iv 6; ib. 9);

for 9 verticals divided over two lines cf. Englund and Nissen

2001:W19408,33.

2. 9 circles (= 90?)

a. No. 24 "reverse" (so called on the expedition

photo): eight of the circles are disposed along the top of the

surface, the ninth below them on the left side of the surface.

This might suggest that the writing began on the left side but it

may be preferable to regard it as an early case of poor planning -

when space ran out after the eighth circle had been impressed, the

writer simply impressed the ninth below the eighth. The "obverse"

of the tablet features 2 circles over three verticals, all

reasonably well centered at the top and in the middle of the

surface respectively. It should be noted that the signs face in a

direction opposite to that of the "obverse," as in No. 18 and in

later Mesopotamian usage (cf. above, IV E). The tablet is very

lightly sealed, if at all.

b. No 26: the tablet is only 4.35 cm wide compared

to 6.5 cm for No. 24, so this time the scribe only managed to

place the first seven circles into the top line, and had to

"indent" the eighth and ninth. But the explanation offered for No.

24 holds, reinforced by the fact that the ninth circle is slightly

23

lower than the eighth.

3. 5 circles over 3 verticals (= 53?)

No. 35: both lines start at the right side of

the tablet, with a large space left between them, as in No. 25,

though the tablet is slightly smaller.

4. 4 circles over 4 verticals (= 44?)

a. No. 22: This is a clear example of its type. It

is impressed on only one surface, and is unsealed. The verticals

are well aligned under the circles, in distinction from the tablet

with the pictogram. The tablet is one of the smallest in the

corpus, the inscribed surface measuring just 2.7x3.9 cm.

b. No. 25: Probably the same type, but this time

the tablet surface is larger (4.25x4.9 cm) and the scribe has

taken advantage of this fact to separate the verticals from the

circles by a considerable blank space, as in No. 35. In the

process, he has failed to align the verticals with the circles.

The tablet is unsealed.

5. 3 circles (= 30?)

a. No. 11: the circles are neatly centered at the

top of the tablet. The rest of the surface and all the other

surfaces are blank.

b. No. 21: three circles, beginning at top right

24

edge, followed by a break and what may be a smaller circle at the

left top edge.

6. 2 circles over 4 verticals (= 24?)

a. No. 14 "reverse" is broken, but the traces

suggest this restoration; alternatively, there may be only three

verticals. At 3.6x5.4 cmm, the tablet is more rectangular than

square, and there is relatively little space between circles and

verticals. The alignment is such that the verticals are not

directly below the circles but to their left (and right). The

"obverse" is inscribed with what appear to be two verticals,

neatly centered at the top of the tablet, as in No, 3. As in No.

18, the tablet has a margin, but only on the "obverse." It is

unsealed.

b. No. 2: the tablet is broken, but may be

restored on the basis of No. 14 reverse, which it resembles in all

respects except that it is sealed.

7. 2 circles over three verticals (= 23?)

See No. 24 "obverse" (above, 1.)

8. 1 circle over 4 verticals (= 14?)

No. 18 "obverse": a well preserved rectangular

tablet, with a deep indentation on the right side and shallower

ones on the left side and top and bottom which may be interpreted

25

as a "ruled margin." A fifth sign is impressed to the left of the

four verticals. One would be inclined to regard this as a crescent

(for the term see below ad No. 12) but for the fact that such

signs otherwise appear on a separate line from the verticals, e.g.

Nos. 4 and 23 rev. The reverse is also inscribed, with a circle

over three verticals, and shows the same kind of margin. In

addition, it has a "horizontal line dividing the field into two

cases" (Schmandt-Besserat 1992:134 and fig. 81).

9. 1 circle over 3 verticals (= 13?)

a. No. 18 "reverse" (see above, 8).

b. No. 3 obverse: there is a slight indentation to

the left of the circle which could be interpreted as a second

circle. However, there seems to be a clear tendency either to

begin the circles on the right hand side of the tablet or to

center the verticals under the circles which seems to be the case

here. There is a margin as in No. 18. The margin is repeated on

the reverse, which features only two verticals, neatly centered at

the top, as in No. 14. Note that they face in the same direction

as the verticals on the obverse, in distinction to No. 18.

10. 1 circle over 2 verticals (= 12?)

No. 9: this is a rather unusual tablet, almost

square (5.3x4.9 cm), though rounded at the top. The circle is

centered at the top of the tablet but the verticals are not

26

centered below it; rather the right one begins under the circle.

11. 1 circle over 1 vertical? (= 11?)

No. 16: this tablet almost gives the impression of

being a practice tablet. Note the atypical placement of the

vertical well to the right of the circle and especially the three

deep circle-like impressions on the short edge.

12. 1 circle over 4 crescents and a pair of joined dots

(= 10.9?)

No. 4 long edge: the tablet is well preserved but

heavily sealed. That may be why the "obverse" is otherwise

uninscribed and the inscription occupies the long edge (as in No.

5). The triangular wedges (so-called by Schmandt-Besserat 1992:135

fig. 84) are here referred to as "fingernail-impressed crescents"

or simply "crescents" following Weiss and Young (1975:88). They

must represent a smaller unit than the vertical wedges, since they

follow these when they co-occur (e.g. No. 23, possibly No. 18). If

(as in later cuneiform) the vertical wedges represent 60 to any

power, including sixty to the power zero = one, then the

triangular wedges may represent a fraction of one, here

hypothetically regarded as 1/5 (see below, VI). They are followed

by an "apex to apex" sign (so called by Schmandt-Besserat 1992:135

fig. 84) or what Weiss and Young call "paired, joined dots"

(1975:88). By the same logic, these are clearly a smaller fraction

27

yet, here assumed to represent 1/2 of a crescent, i.e. 1/10 of a

vertical (see below, VI).

The reverse of the tablet has four verticals, neatly

centered at its top.

13. 1 circle over 2 crescents (= 10.4?)

No. 30: the amount is just fractionally less than

No. 4 (above, 12).

14. 9(?) verticals (= 9?)

No. 10: the tablet is damaged but space

considerations suggest room for up to three verticals at the top

right for a total of nine verticals.

15. 5 verticals over 2 crescents over 1 pair of dots

(5.5?)

a. No. 23 "obverse": the verticals may be said to

be centered, leaving a slight space on the right. In this space

there are the merest traces of an incised sign, vaguely resembling

a hand. The crescents are definitely centered, being placed to

right and left of the middle vertical. The pair of dots is near

the right edge. The "reverse" has three verticals over 2 crescents

(= 3.4?). The tablet is unsealed.

b. No. 36: the upper right hand corner of the

obverse is lost, so the restoration is speculative, based on the

28

assumption that the verticals began near the right edge. The

reverse probably had four verticals, as in No. 4, which this text

resembles, except that it is not sealed.

16. 4 verticals (= 4?)

a. No. 5: the inscription is on the long edge; the

tablet is sealed.

b. see No. 4 reverse (above 12).

c. see No. 36 reverse (above 15b).

17. 3 verticals and 2 crescents (= 3.4?)

No. 23 "reverse": see above, 15a.

18. 3 verticals (= 3?)

a. No. 20: the verticals are fairly well centered

at the top of the tablet, which is sealed. On the drawing, the

reverse shows four crescents; note that they face in the opposite

direction from the verticals on the "obverse."

b. No. 28: the inscribed side is damaged, but

otherwise resembles No. 20, except that it s unsealed.

19. 2 verticals over 4 crescents and 1 pair of dots (=

2.9?)

No. 31: Note that the pair of dots is on the same

line as the crescents (so too in Nos. 4 and 29); the fact that it

29

is to the left of the crescents is another powerful argument for

reading the lines from right to left. The long edge has two

crescents.

20. 2 verticals over 2 crescents (= 2.4?)

No. 34.

21. 1 vertical over 2 crescents and 1 pair of dots (=

1.5?)

No. 29: Unsealed.

22. 4 crescents (0.8?)

No. 20 "rev." (see above, 18a).

23. 2 crescents (0.4?)

No. 31, edge (see above, 19)

C. Fragmentary tablets with impressed signs

1. No. 1: fragment with traces of a circle over a

double axis.

(No. 2: fragment from left side: 2 circles over three

verticals; for suggested restoration see above, ....)

2. No. 6: fragment from right side: 2 verticals.

3. No. 7: fragment from bottom(?): 2 verticals.

30

4. No. 8: fragment from upper right corner: 2 verticals

preserved.

5. No. 12: fragment from right side(?): traces of 2

signs

6. No. 13: fragment from left side: 2 circles.

7. No. 15: fragment from left side: 2 circles.

8. No. 17: fragment from left side, remains of 1 circle

at top.

9. No. 19: fragment from left bottom: 2 verticals.

10.. No. 32: fragment from bottom (?), traces of 2 (?)

signs..

11. No. 37: fragment from right side, to judge by the

orientation of the large vertical impressed on the short edge (cf.

No. 33, above 1).

12. No. 38: fragment from top: two circles and remains

of a third.

VI. Discussion

A. The commodity or material being counted

The token system that preceded the number-tablets

indicated not only the quantity but the nature of the commodities

being counted, according to Schmandt-Besserat. The slightly later

archaic tablets from Uruk IV perpetuate these distinctions, albeit

on a more sophisticated level, using different systems of counting

31

for different commodities. Since the Godin system most nearly

resembles that used for barley at Uruk, one may venture a guess

that barley is also at issue here (see below, VII). The Godin V

stratum where the tablets were found indeed contained storage

rooms and pottery (Young 1997) and there were "legumes and grains

stored in room 22" (Weiss and Young 1975:96). The tablets, though

found in other rooms and in the courtyard, may refer to the grain

stored in the storage in room 22.

B. The possibility of pairs. Note that there may be

several pairs of duplicate or near duplicate texts, notably Nos.

22 and 25 and Nos. 5 and 21, all unsealed, and inscribed on one

side only. In some cases, one duplicate is sealed and the other is

not, namely Nos. 2 (sealed) and 14 (unsealed) and Nos. 20 (sealed)

and 28 (unsealed); again, all four are inscribed on one side only.

Nos. 24 (sealed ??) and 26 (unsealed) form a similar pair. Where

more than one side is inscribed, only one duplicates one or even

two other tablets, namely: Nos. 3 rev. and 18 rev., Nos. 23 and 36

obv. (all unsealed), and Nos. 4 and 5 (both sealed) and 36 obv.

(unsealed). Whether any significance can be attached to these

observations remains to be seen.

C. The possibility of joins. While it is difficult to

propose or even to rule out joins on the basis of photographs and

drawings alone, there are no likely prospects for joins among the

fragments. CUYLER: TRY IT ON BASIS OF MY DESCRIPTIONS OF THE

FRAGMENTS??

32

VII. The numbers

Since the inscribed Godin tablets consist almost

entirely of numbers, an attempt to assess the relative and

absolute size of these numbers is called for. "Number tablets"

from other sites need to be considered, but it must be remembered

that there is not as yet a conventional form of notation that

bridges the geographical horizon of all the finds, each of which

has its own divergences; but even when there are trans-regional

similarities in the number signs, there may not be any easy way to

compare them with the forms of numeral notation in historic

periods (Nissen, Damerow and Englund 1990:172f. = 1993:128f.).

A. Given the local variations in notation, and the

uncertainties attached to all of the local varieties, the best

that can be done for now is to advance a working hypothesis as to

both relative and absolute numerical values to be attached to the

signs in the Godin tablets. In a limited corpus, it is impossible

to differentiate among possibly multivalent signs; hence it is

best to operate with the further hypothesis that whatever was

being counted or measured was one and the same staple, or at least

used one and the same numeration. On this basis the safest course

is to regard each lower unit as related to the next higher unit in

the ratio of x+1:1 where x = the highest number of units attested

for the lower unit.

33

TABLE

large circle vertical crescent pair value

vertical of dots

A 27 3 3 33

B 1 33 2 4 6 246

2 24,26 9 90

3 35 5 3 53

4 22, 25 4 4 44

5 11, 21 3 30

6 2, 14 2 4 24

7 24 2 3 23

8 18 1 4 14

9 3, 18 1 3 13

10 9 1 2 12

11 16 1 1 11

12 4 1 4 1 10.9

13 30 1 2 10.4

14 10 9? 9

15 23, 36 5 2 1 5.5

16 4, 5, 36 4 4

17 23 3 2 3.4

18 20, 28 3 3

19 31 2 4 1 2.9

34

20 34 2 2 2.4

21 29 1 2 1 1.5

22. 20 4 0.8

23. 31 2 0.4

On this basis, and without reference for now to archaic

tablets from other sites, it will here be posited that the "pair

of dots" sign, which occurs five times, each time by itself,

stands in the proportion of 1:2 to the crescent. The latter occurs

9 times, six times as two of the kind and three times as four of

the kind. If four is in fact the highest number of crescents to

occur in the texts, then it may tentatively be suggested that it

stands in the proportion 1:5 to the vertical. The latter occurs no

less than 25 times, in quantities ranging from 1 to 9. It would

then be likely to relate to the circle as 1:10. The circle occurs

19 times, in quantities ranging - again - from 1 to 9. It should

thus relate to the large vertical as 1:10. Therewith a basically

decimal system has been reconstructed, although the lowest unit is

subdivided in two.

When it comes to absolute size, one is forced - again

ex hypothese - to fall back on the known value of Mesopotamian

number signs in the sexagesimal system. There the vertical wedge

is universally employed to represent 60 to any power, including 60

to the power zero = 1. Applying that value to the vertical at

35

Godin, we can hypothetically assign to the successive signs in

descending order the values 100, 10, 1, 1/5 and 1/10. The values

assigned to the numbers in the texts above are based on this

understanding.

B. Base 10 ?

An apparent exception to the observation that no more

than nine circles appear on any one tablet is found at various

sites. At Jebel Aruda, one small tablet has two lines of five

circles each, and a large tablet has two lines of eleven circles

each (Nissen, Damerow and Englund 1990:174, fig 18h = 1993:130,

fig. 114). At Sialk, one tablet has two lines of 10 circles each

starting at the top right edge, followed by a third line with

three circles. This tablet further diverges from the norm at Susa

and Godin by starting the second and third lines before the

preceding line is carried all the way to the left edge (Ghirshman

1938:67f. and pl. xcii S. 1631). Perhaps it represents some kind

of an exercise tablet. Note that its "reverse" has an intricate

drawing much like some of the school-texts from Ur and Shuruppak

(Fara; Heinrich 1931:64). In and of itself it is not enough to

negate the theory of a decimal base for counting. On the other

hand there seems no doubt that a decimal base was early replaced

by a sexagesimal base as illustrated by 'the oldest known

schooltext from Babylonia' (Englund and Nissen 2001:17 ad W

19408,76+) in which 600 is expressed by a ligature of 60 and 10

(Nissen, Damerow and Englund 1990:98 = 1993:58). It is

36

dramatically illustrated in a Fara period text which is described

as "the oldest document of its kind" and as one of "the oldest

documents relating to a systematic study of the arithmetical

techniques which originated in the archaic period" (Nissen,

Damerow and Englund 1990:181-3 and fig. 18m = 1993:136-8 and fig.

119).

Another possible exception may be represented by

the lone tablet from Mari (Parrot 1965:12), with 4 circles in the

first line, 6 in the second, and one in the third, followed by

three verticals. In its isolation, it can hardly be treated as

normative.

There is also the question of markings which are

neither as deep as the verticals nor as rounded as the crescents

though, like the latter conceivably made with the fingernail. The

presence of five of these on an Uruk tablet (UVB 25 pl. 27k,n) may

have to be explained as representing something other than a

number.

C. Large vertical or large circle?

For No..33 (and possibly No. 37) we have

interpreted the large impressions as large verticals, assigning

them a value of 100. But it must be admitted that there are clear

examples of tablets where a large circle is written over one or

more (small) circles, e.g. W 20239 in Englund and Nissen 2001. At

Jebel Aruda, both phenomena occur (van Driel 1982:16).

D. Fractions

37

To interpret the vertical wedge as "1" is to be

guided not only by later cuneiform usage but by the virtually

universal meaning of this sign. In treating the units that follow

this sign as fractional values of "1," the later analogies are

more ambivalent. The archaic roster of numeral signs that precedes

full cuneiform writing, or "proto-cuneiform," counts no less than

60 different forms according to the tabulation by Nissen, Damerow

and Englund (1990:62, fig. 8a = 1993:26, fig. 27), but not one of

them resembles a crescent scratched with the fingernail. The

nearest analogue is their N 39 (= Falkenstein 1936, Sign No. 935)

which, however, faces the other way: if the Godin sign may be

compared to a growing crescent, the Uruk sign is a waning

crescent. It is related to the vertical wedge, which is the next

highest unit in this system, as 1:5, just as (on my reading) the

crescent at Godin, and presumably on the same basis, i.e. that it

never occurs more than four times at one time. It is used in the

basic barley-system ("SE-system"), and recurs, with an additional

stroke, in the special barley system apparently intended for malt.

While the value of the crescent as 1/5 had been

recognized already by Langdon (1928) and Falkenstein (1936), the

value of the next lower unit (Nissen et al. N 24 = Falkenstein

1936, Sign No. 919) was first established by Vaiman, who

recognized the sign at Uruk resembling the "pair of dots" at Godin

as 1/2 of 1/5, i.e. 1/10. Essentially the sign signifies a

38

reciprocal of 2. Similarly, three dots represent the reciprocal of

3, i.e. 1/3, more particularly 1/3 of 1/5 = 1/15; four dots

represent 1/4 of 1/5 = 1/20, 5 dots 1/5 of 1/5 = 1/25, and 6 dots

1/6 of 1/5 = 1/30 (Vaiman 1976:18f.).

These smaller fractions occur in the barley-system at

Uruk (Nissen et al. N 26, 28, 29, 30a). In the malt-system at Uruk

the paired dot is said to be in the ratio of 1:5 to the crescent,

hence of 1:25 to the vertical, and for good measure in the "EN-

system" (perhaps used for measures of weight) in the ratio of 1:2

to the vertical (without any intervening crescent unit). The sign

N 51 also bears some comparison, but is not explained at all

(Nissen, Damerow and Englund 1990:64f., fig. 8b = 1993:28f., fig.

28). At Godin, by contrast, the paired dot occurs on five separate

tablets but never more than once at a time, i.e. only in the basic

meaning of 1/2 of 1/5 = 1/10.

The fact that the form, if not always the function, of

the number-signs on the Godin tablets most nearly resembles the

form of those reconstructed for the tablets from Uruk for the

measurement of barley (both below and above "1") may provide a

precious clue to the identity of the staple being accounted for at

Godin (above, VI A).

Comparing Godin-type tablets from sites other than Uruk

yields no comparable signs for fractions. It is interesting,

however, that the paired dots occur in proto-Elamite, and likewise

with the value 1/2 (Meriggi 1971:167, 169), while the (growing)

39

crescent occurs at Jemdet Nasr, apparently with the value 1/5,

i.e. it "never occurs more than four times" (Langdon 1928:65, No.

451).

VII. Conclusion: regional and inter-regional implications

A. Implications for Proto-Elamite

The relations of the Godin tablets to proto-Elamite

deserves closer scrutiny. True, "if the people who occupied the

Oval Enclosure [at Godin] were merchants, they might just as well

have been merchants from Sumer as from Susa" (Young 1986:218). But

given the appearance of Godin-type number tablets not only at the

type site but at four other sites to its south, all or most of

them in the highland kingdom later known as Elam, it is worth

seeing whether they have any relationship to the proto-Elamite

tablets known from Sialk, Shahr-i Sokhta and particularly from

Susa. Tablets inscribed with this script first appear at Susa in

Level Cb, which is said to correspond to Uruk III. The

conventional wisdom therefore has it that this script was derived

from the first full script attested at Uruk in Level IV (e.g. Hinz

1973:28). But others argue that it goes back to the middle of the

fourth millennium and that it originated and developed

independently of Sumerian cuneiform and its forerunners (Gragg

1995:2162f.). So it may be worth noting certain resemblances

between the Godin-type tablets and these earliest proto-Elamite

40

tablets, resemblances that are not found in the archaic tablets

from Uruk.

In the first place, the disposition of the number signs

on the tablet are such that they begin at the top right corner of

the tablet (interpreting "top" as in the Godin tablets); the

earliest Uruk tablets tend to dispose the signs much more often

symmetrically in the middle of the top of the tablet. If the

nature of the commodity counted is identified, the pictogram or

logogram in question appears before, i.e. to the right of the

number signs, as in Godin No. 27. Both these features are

illustrated on a tablet published by Walker, who also discusses

the directions in which the signs were inscribed and read

respectively (1980 75 and fig. 4). Other examples, also probably

from Susa, were published by Stolper (1978 figs. 15, 17, 18).

Secondly, the proto-Elamite tablets favor the Godin

type of "growing crescent" (sometimes followed by a single pair of

dots) against the "waning crescent" in use at Uruk (e.g. Stolper

1978:96 fig 17).

Thirdly, the numbering system in proto-Elamite appears

to be decimal rather than sexagesimal, in this regard, too,

resembling Godin more than Uruk (Gelb 1952:89; Brice 1962; Meriggi

1971:159-172).

These are no more than hints, but they should suggest

the need to reconsider the precise route which led to proto-

Elamite writing. Again, it is necessary to put a limit on

41

speculation, however. Thus the suggestions of a proto-Elamite or

Elamite connection to Dravidian, however beguiling, cannot be

pursued here (McAlpin 1974, 1975).

B. Implications for "Greater Mesopotamia"

In spite of the near uniqueness of the Godin

corpus in respect to fractions of "1," it is linked to archaic

number tablets from elsewhere in all other characteristics. In

fact, the similarities are striking, the divergences minor, when

one considers such factors as the date of the assemblage, the size

and shape of the tablets, the appearance of the units from "1" and

up, the arrangement of the number signs on the tablets, the use of

seal impressions, and of course the medium (clay) on which the

seals and number tokens are impressed. Clearly, the Godin material

is part and parcel of an inter-regional phenomenon that embraces,

in modern terms, Iraq, northern Syria and western Iran or, in

ancient terms, what has come to be regarded as "Greater

Mesopotamia." The term was introduced by Flannery (1965:1247) and

the concept was utilized by Young in his discussion of early

population dynamics (1972:828, 1977:828; Smith and Young 1972:3),

becoming the theme of his tribute to Robert Dyson (Levine and

Young 1977; cf. also Weiss and Young 1975:94; Hallo 1980:309 and

n. 8). It is a useful reminder of what others have called "the

Uruk expansion" (e.g. Algaze 1993:53-55, 98). At the same time it

helps to define the limits of that expansion. Thus it was easy to

exclude the material from Shahr-i Sokhta, Pakistan, Romania and

Russia from the discussion (above, II).

(8357 words)

40 pp.

1. His covering letter to me, dated June 7, 1974. The memo itself carried the date of April 16, 1974.

Godin-Bibliography Algaze, Guillermo, 1993: The Uruk World System: the Dynamics of Expansion of Early Mesopotamian Civilization (Chicago/London: University of Chicago). Amiet, Pierre and Maurizio Tosi, 1978:"Phase 10 at Shahr-i Sokhta: Excavations in Square XDV and Late 4th Millennium B.C. Assemblage at Sistan, with an epigraphic note by Pero Meriggi," East and West 28:9-31 and 35 figs. Anonymous, 1977: "C'etait ecrit: des pictogrammes de plus de 5.000 ans dnord-ouest du Pakistan," Le Figaro, 18 August 1977:14. Badler, Virginia R., 19XX: Godin V and the Chronology of the Late Uruk Period (Ph.D. Thesis, University of Toronto). Brice, W.C., 1962: "The Writing System of the Proto-Elamite Account Tablets of Susa," BJRL 45:15-39. Collon, Dominique and Julian Reade, 1983: "Archaic Nineveh," BaM 14:33-41. Curtis, John, ed., Fifty Years of Mesopotamian Discovery (London: British School of Archaeology in Iraq). Deimel, Anton, 1928: Sumerisches Lexikon II. Teil Band 1 (Rome: Pontificum Istitutum Biblicum). Delougaz, P., 1972: Memorial Volume of the 5th Intl. Congress of Iranian art and archaeology (Teheran).

42

Edzard, Dietz O., 1980: "Keilschrift," RLA 5:544-568. Englund, Robert K. and Hans J. Nissen, 2001: Archaische Verwaltungstexte aus Uruk: die Heidelberger Sammlung (Berlin: Mann). Falkenstein, Adam, 1936: Archaische Texte aus Uruk (Berlin: Harrassowitz). Fischer, Claudia, 1997: ""Siegelabrollungen im British Museum auf Ur III-zeitlichen Texten aud ser Provinz Lagas," BaM 28:97-183. Flannery, K.V., 1965: "The Ecology of Early Food Production in Mesopotamia," Science 147:1247-1256. Gelb, I.J., 1952: A Study of Writing: the Foundations of Grammatology (London: Routledge and Kegan Paul). Ghirshman, Roman, 1938: Fouilles de Sialk I (Paris: Geuthner). Glassner, Jean-Jacques, 2000: Ecrire a Sumer ([Paris]:Editions du Seuil). Gragg, Gene B., 1995: "Less-Understood Languages of Ancient Western Asia," in Civilizations of the Ancient Near East (4 vols.), ed. Jack M. Sasson (New York: Scribner's), IV:2161-2179. Green, Margaret W. and Hans J. Nissen, 1987: Die Zeichenliste der archaischen Texte aus Uruk (= ADFUW 11) (Berlin: Gebr. Mann). Hallo, William W., 1980: "The Expansion of Cuneiform Literature," Proceedings of the American Academy for Jewish Research 46-47:307-322. _____, 1982: review of C.B.F. Walker, Cuneiform Brick Inscriptions in JCS 34:112-117. _____, 1996: Origins (Leiden etc.: Brill) Hansman, J., 1978: "The Question of Aratta," JNES 37:331-336. Heinrich, Ernst, 1931: Fara (Berlin: Staatliche Museen). Hinz, Walther, 1973: The Lost World of Elam (New York: New York University), tr. by Jennifer Barnes from Das Reich Elam (Stuttgart:Kohlhammer, 1964).

43

Labat, Rene, 1948: Manuel d'Epigraphie Akkadienne (Paris: Geuthner), rev. ed. 1976. Langdon, Stephen, 1928: The Herbert Weld Collection in the Ashmolean Museum: Pictographic Inscriptions from Jemdet Nasr (OECT 7) (London etc.: Oxford U.P.). LeBrun, Alain and Francois Vallat, 1978: "L'origine de l'ecriture a Suse," DAFI 8:11-59. Levine, Louis D. and T. Cuyler Young, Jr., 1977: Mountains and Lowlands: Essays in the Archaeology of Greater Mesopotamia (= BM 7) (Malibu: Undena). Majidzadeh, Yousef, 1976: "The Land of Aratta," JNES 35:105-113. Marshack, Alexander, 1979: "Upper Paleolithic Symbol Systems of the Russian Plain: Cognitive and Comparative Analysis," Current Anthropology 20:271-311. Matthews, Donald M., 1997: The Early Glyptic of Tell Brak: Cylinder Seals in Third Millennium Syria (Orbis Biblicus et Orientalis. Seria Archaeologica 15) (Gottingen: Vandenhoeck & Ruprecht). ________, and Jesper Eidem, 1993: "Tell Brak and Nagar," Iraq 55:201-207. McAlpin, David, 1974: "Toward Proto-Elamo-Dravidian," Language 50:89-101. _______, 1975: "Elamite and Dravidian: Further evidence of relationship," Current Anthropology 16/1:105-116. Meriggi, Piero, 1971: La Scrittura Proto-Elamica I: La scrittura e il contenuto dei testi (Rome: Accademia Nazionale del Lincei Michalowski, Piotr, 1986: "Mental Maps and Ideology: Reflections on Subartu," in The Origins of Cities in Dry-Farming Syria ..., ed. Harvey Weiss (Guilford, CT: Four Quarters) 129-156. Nissen, Hans J., 1977: "Aspects of the Development of Early Cylinder Seals," in Seals and Sealing in the Ancient Near East, ed. by McGuire Gibson and Robert D. Biggs (= BM 6) (Malibu: Undena) 15-23.

44

______, Peter Damerow and Robert K. Englund, 1990, 2nd ed. 1991: Fruhe Schrift und Techniken der Wirtschaftsverwaltung im alten Vorderen Orient (Berlin?: Franzbecker). Translated as: ___________________________________________, 1993: Archaic Bookkeeping: Early Writing and Techniques of Economic Administration in the Ancient Near East, tr. by P. Larsen (Chicago: University of Chicago Press). Oates, David, 1982: "Tell Brak" in Curtis 1982:62-71. _____, Joan Oates and Helen McDonald, 2001: Excavations at Tell Brak (vol. 2) (London: British School of Archaeology in Iraq). Parrot, Andre, 1965: "Les Fouilles de Mari: quatorzieme campagne (Printemps 1964)," Syria 42:1-24 and pls. i-ii. Porada, Edith, 1965: "The Relative Chronology of Mesopotamia, Part I," in Robert W. Ehrich, ed., Chronologies in Old World Archaeology (Chicago/London, Univesrity of Chicago Press), 133-200. ______, et al., 1992: "The Chronology of Mesopotamia, ca. 7000-1600 B.C.," in Robert W. Ehrich, ed., Chronologies in Old World Archaeology (3rd ed., 2 vols., Chicago/London: University of Chicago), I 77-121, II 90-124. Reade, Julian, 1992: "An Early Warka Tablet," in Von Uruk nach Tuttul: eine Festschrift fur Eva Strommenger, ed. Barthel Hrouda et al. (Munich/Vienna: Profil) 177-179 and pl. 79. Salonen, Armas, 1966: Die Hausgerate der alten Mesopotamier (= AASF B 144) (Helsinki, Suomalainen Tiedeakatemia). Scheil, Vincent, 1923: Textes de Comptabilite Proto-Elamites (nouvelle serie) (= MDP 17) (Paris: Leroux). Schmandt-Besserat, Denise, 1981: "Decipherment of the Earliest Tablets," Science 211:283-285. _________________, 1992: Before Writing I: From Counting to Cuneiform, foreword by William W. Hallo (Austin: University of Texas). _________________, 1992a: Before Writing II: A Catalog of Near Eastern Tokens (Austin, Uiversity of Texas).

45

Smith, Philip E.L., and Young, T. Cuyler, Jr., 1972: "The Evolution of Early Agriculture and Culture in Greater Mesopotamia: a Trial Model," in PopulationGrowth: Anthropological Implications, ed. Brian Spooner (Cambridge, MIT) XXXXXXX. Stolper, Matthew W., 1978: "Inscribed Fragments from Khuzistan," DAFI 8:89-96. Strommenger, Eva, 1973: "Ausgrabungen in Habuba Kabira und Mumbaqat," AfO 24:168-171. ___________, 1977: "Habuba Kabira am syrischen Euphrat: Grabungen der DOG 1969-1975," Antike Welt 8/1:11-20. __________, 1981: Habuba Kabira: eine Stadt vor 5000 Jahren (Mainz: von Zabern). Vaiman, A.A., 1976: Uber die protosumerische Schrift," in Wirtschaft und Gesellschaft im alten Vorderasien, ed. J. Harmatta and G. Komoroczy (reprinted from AAASH 22, 1974), 15-27. Vallat, F., 1971: "Les documents epigraphiques de l'Acropole," DAFI 1:235-245, figs. 31-72, and pls. xxivf. ______, 1973: "Les tablettes proto-elamites de l'Acropole (Campagne 1972)," DAFI 3:93-103 ______, 1979: "Le materiel epigraphique des couches 18 a 14 de l'acropole," in La Sequence archeologique de Suse ... ed. J. Perrot (= Paleorient 4, 1978:131-228) 193-195. Van der Leeuw, S.E., 1972: "Jebel Aruda," in Exposition des Decouvertes de la Campagne Internationale de sauvegarde des antiquites de l'Euphrate (Musee National d'Alep, ed. Direction Generale des Antiquites et de Musees de la R.A.S. (XXXXXX), XXXXX Van Driel, G., 1982: "Tablets from Jebel Aruda," in ZIKIR SUMIM: Assyriological Studies Presented to F.R. Kraus, ed. G. van Driel et al. (Leiden: Brill) 12-25. Vlassa, N., 1963: "Chronology of the Neolithic in Transylvania, in the light of the Tartaria Settlement's Stratigraphy," Dacia n.s. 7:1-10 + 1 fold-out. Walker, C.B.F., 1980:"Elamite Inscriptions in the British Museum," Iran 18:75-81 and pls. if.

46

Weiss, Harvey and T. Cuyler Young, Jr., 1975: "The Merchants of Susa: Godin V and Plateau-Lpwland Relations in the Late Fourth Millennium B.C.," Iran 13:81-97. Young, Jr., T. Cuyler, 1972: "Population Densities and Early Mesopotamian Urbanism," in Man, Sttlement and Urbanism, ed. P.J. Ucko et al. (London: Duckworth) 827-842. _____, 1974: "The Day We Found the Tablet," Rotunda 7/2:4-11. _____, 1977: "Population Dynamics and Philosophical Dichotomies," in Levine and Young 1977:387-398. _____, 1983: "Godin Tepe Period VI/V and Central Western Iran at the End of the Fourth Millennium," in Gamdat Nasr: Period or regional Style?, ed. Uwe Finkbeiner and Wolfgang Rollig (TAVO Beihefte B 62) (Wiesbaden, Reichert) 212-228. _____, 1997: "Godin Tepe," Oxford Encyclopedia of Archaeology in the Near East, ed. Eric M. Meyers (New York/Oxford: Oxford U.P., 1997) vol. 2:416f. ABBREVIATIONS Studies Kraus

Godin appendix After competing the draft of my chapter on the Godin Tepe tablets, sent copies to a number of colleagues who really iknow the subject ofarchaic tablets. Two of them were kind enough to resond with.very thoughtful comments. By rights I should have incorporated all of these into my draft, aand did so where feasible But most of them are based on premises so different from my own that it would have been impossible to integrate their explanations with mine. With their consent, I have instead opted to append their comments to my chapter under their own names and in their own words. I am deeply grateful to them, both for their permission and for their critiques. The following are virtually verbatim transscipts of their remarks. My editorial comments are kept to a minimm, and idenified by square brackets.

47

I. Robert K. Englund, letter (e-mail) of May9, 2003:

Dear Bill, Now then, I’ve gone through the Godin Tepe article and recommend that you review the comments on the numerical notations on the texts. Please note that there is no decimal system on any Late Uruk material, thus for instance your text 33 must be a simple exagesimal notation, whereas the textx 24 and 25 that you cite there shoul be grain notations thazt, assuming both columns are to be read together, represent 9x6 = 54 basic units, This might also be simple doodling of [on] the reverse surfaces of both texts, I should think that nearly all of your accounts consist of sucg grain nottions, but this could be contested in those cases thz contain”N1” and/or “N14”in numbers that could be interpreted as sexagesimal. All notations with “N39b”, “N24”must be from [the] grain capacity system. N39b (“waning crescent”) is common in the Uruk Iva textx, replaced in Uruk III by N39a. Falkenstein 1936 should not be cited inztead of ATU 2 (there page 342). The value of N39 (1/5 NI) and lower mrembers of the grsain system was clearly demonsrated by Falkenstein “Archaische Texte des Iraq-Museums in Bagdad”, in OLZ 40,[(1937)] pp. 401-410, and is also clear from totals Peter [Damarov] and I discussed in ATU 2 ch.3. I do not understand your comments about the malt system in VIImD. It mirrors the basic system. The proto-Elamite decimalsystem is not attesdted at this time; the Godin texts are all Late Uruk. Some notes: Text 10: if 6N1 then not grain notation. Text 13: is that a lsarge N14 on obverse? If so, then the notsation could be sexagesimal or grain (or bisexagesimal), Text 27: the famous one, this could be either sezagesimal “33” or a grain notation equal to 3x6 + 3 = 21 N1. Text 40: (and perhaps one or severalthat I cannot recognizew, nos. 34, 37 too?) looks to have a “60”sign on the edge; this is proably in ideographic use according to a nmber of proto-Elamite texts. Text 34 ([10??]}: as I mentiones, in all likelihood sexagrsimal and therefore very interesting here, This cannot be grain since 6N1 would have been replaced by N14. Other notes working from beginning of [chapter]: Van Ess of DAI-Warka dates Uruk Iva now to 3350, and D. Suerenhagen in his Habil[itationsschrift]mdates the Red Temple “Uruk IVa” tests toat least Ivb, and probably V. Mathematicians take exception to the use of “numerals” to

48

describe “numerical signs” in the archaic texts. The counters matter should be referenced strictly to Lieberman’s critical review of [Schmandt-Besserat], “Of Clay Pebbles, Hollow Clay Balls, and Writing: a Sumerian View” in: AJA 84 [(1980)], pp. 339-358. You know my and others’ belief that there is no clear evidence for the complex token theory, and much that speaks agaionstit. The Uruk “tablets” (unclear function) W 10133 (six tablets) and 16184 (and all other gypsummtablets) are published in ATU 5. Mathews wrote in MSVO 2 of the Jemdetr Nasr practice of sealing first, and writing over the impression [of the seal(s)].He consideras this a “letter head”. If you transliterated the texts using the Green-Nissen sign list [Margaret W.Green and Hans J. Nissen, Zeichenliste der archaischen Texte aus Uruk (Berlin, Mann, 1987)], instead of describi ng the notation with circles and verticals, it would help to avoid confusion with the standard archaic publications that will describe N1 as “horizontal” and so on – the old orientation flap. A “vertical”, in our pulications = ½ of the basic unit in the sexagesimal system.

II. Denise Schmandt-Besserat, letter of March 29, 2003 Dear Bill , Thank you for sharing with me your ms. …I hope the attached will be of some help. Joran Friberg might be recognized as having done seminal work on

the archaic tablets, namely interpreting the impressed signs as units of commodities – not numerals. According to him, wedges + circular signs are units of cereals.See Joram Friberg, The Third Millennium [Early]Roots of Babylonian Mathematics (Chalmers University of Technology and theUniversxity of Goteborg, 1978-79), pp.. 10, 21 and 46. I have followed Feiberg’s interpretation (see my Before Writing, 1992, pp.1-132) and so have Nissen, Englund and Damerow (see their Archaic Bookkeepimg, 1993, pp. 28-29) previously eidem, Fruhe Schrift und Techniken der Wirtschftsverwaltung im altren Vorderen Orient (1990, pp. XXX)].

[Comments to selected paasages :] II.I think it is better to call the hollow objects holding tokens

“envelopes” and reserve the term “bulla” for the solid

49

50

objects only (to string tokens or later to identify scrolls)..

For the sake of forging aseries of standard terms, Iwould suggest

to [replace]the term “Godin-type tablet” by “impressed tablet”.

“Other Iranian sites that have yielded archaic tablets” [add:] Tepe Yahya and Malyan: [for

which] see: Peter Damerow and Robert Englund, The Proto-ElamiteTexts from Tepe

Yahya (Cambridge, MA: Peabody Museum of Archaeology and Ethnology, 1989) =

American School of Prehstoric Research Bulletn, vol. 39; [respetively:] Matthew W.

Stolper, “Three Stray Elamite Tablets freom Malyan,” in Naomi F. Miller and Kamyar

Abdi, eds., Kedi Bud, Yeki Nabud: Essays on Archaeology of Iran in Honor of William

M. Summer, = Monograph 48, Cotsen Institue of Archaeology, University of California

at Los Angeles, 2003, ch. 17. III. “Clay tokens such as occur as early as the ninth millennium”: th first tokens occur ca. 7500 BC (at Mureybat, Syria, and Tepe Asiab, Iran), therefore: “eighth millenium”.